| Abstract: | Despite the general success of species envelope models, capturing the fine-scale detail of patchiness in the distributions of some species is problematic. For great bustards in Spain, apparently suitable habitat patches remain unoccupied and cannot be distinguished from occupied patches in current distribution models. We consider philopatry and conspecific attraction as main behavioural mechanisms which could account for this patchiness, and then look for evidence of their influence on the distribution of great bustards across the whole of Spain. We compared the characteristics of habitat patches classed as suitable by a distribution model according to whether they were actually used or not. Occupied patches were larger than unoccupied patches and over-used in proportion to their size, suggesting aggregation and a metapopulation structure. Arguing that conspecific attraction may serve to transfer information about site history and environmental predictability (at least over a short time period), we compared the coefficients of variation in time-series of vegetation and climatic factors at occupied and unoccupied sites. Great bustards chose sites which were more environmentally stable at critical periods in the breeding cycle, "public information" that can only be gained from others rather than through sampling. There is thus evidence that both metapopulation dynamics and conspecific attraction influence the large scale distribution of great bustards in Spain. We discuss how alternative predictor variables and multi-stage analyses may help us to incorporate behavioural mechanisms into distribution models, but acknowledge that there are limits to the value of species envelope models for animal species making decisions. |
